How to Define a North Star Metric for Your Product

Choosing a North Star metric isn't about picking an inspiring-sounding number or copying what a famous company like Airbnb or Facebook uses — it's a specific, structured process of identifying the one number that genuinely reflects the real value your product delivers, and that reliably predicts long-term business success for your specific situation. This guide walks through that process step by step.

Quick facts

• A North Star metric should reflect real customer value delivered, not just company revenue or vanity activity.
• The right metric is specific to your product — copying another company's North Star metric without validating it fits your situation is a common mistake.
• The process involves identifying core value, finding a measurable proxy, and validating it against real business outcomes.
• See the conceptual framework and famous examples in The North Star Metric Framework Explained.

Step-by-step process to define your North Star metric

1. Write down, in plain language, the core value your product actually delivers. Not a feature list — the underlying outcome customers get. What are they able to do or achieve because of your product that they couldn't do (as easily) without it?
2. Brainstorm several possible metrics that could serve as a proxy for that value. Don't settle on the first idea — generate a genuine range of candidates before narrowing down.
3. Check each candidate against a simple filter: Could this number go up in a way that doesn't reflect real customer value? If yes, it's a weak candidate — this filters out vanity-metric-style options.
4. Validate the strongest candidate against real historical data, if available. Does this metric actually correlate with retention, revenue, or other confirmed business outcomes in your own data? A theoretically sound metric that doesn't hold up against your real data is a weaker choice than one that does.
5. Confirm the metric is simple enough for the whole organization to understand and rally around. An overly complex, multi-part formula defeats the purpose of a single, shared North Star.
6. Roll it out deliberately, explaining the reasoning behind the choice, not just announcing the number — teams need to understand why this metric was chosen to genuinely rally around it, not just what it is.

A worked example of the process

A team building a recipe and meal-planning app works through this process:

Step 1 — Core value: Helping people cook satisfying meals at home more easily and with less daily decision fatigue.

Step 2 — Candidate metrics: Total app downloads, weekly active users, recipes saved, meals actually cooked (self-reported or inferred from behavior), subscription revenue.

Step 3 — Filtering: Downloads can rise through marketing spend without reflecting real ongoing value — filtered out. Weekly active users is closer, but someone could open the app without genuinely using it meaningfully — a weaker signal. "Meals actually cooked" ties most directly to the core value identified in Step 1.

Step 4 — Validation: The team checks historical data and finds that users who cook 3+ meals per week using the app have dramatically higher 90-day retention than users who don't — confirming this metric genuinely predicts long-term business success, not just sounding reasonable in theory.

Step 5-6: "Meals cooked per week" becomes the chosen North Star metric, rolled out with clear reasoning tied to both the product's core value and validated retention data.

Why validating against real data matters so much

It's tempting to choose a North Star metric based purely on how well it sounds conceptually, without checking whether it actually correlates with real business outcomes in your own data. A metric that seems intuitively right can still fail this test — validating against real historical data (do users who score well on this metric actually retain better, spend more, or refer others more) is what separates a genuinely strong North Star metric choice from one that just sounds good in a strategy meeting.

What to do if you don't yet have enough data to validate

Early-stage products often don't have enough historical data yet to fully validate a North Star metric candidate. In this situation, it's reasonable to choose a metric based on the strongest available reasoning from steps 1-3, while treating it explicitly as a working hypothesis to be revisited and validated once more data accumulates — rather than either avoiding choosing one at all, or treating an unvalidated choice as permanently settled.

Common mistakes when defining a North Star metric

• Copying a well-known company's North Star metric without checking if it actually fits your product. What worked for Airbnb or Spotify reflects their specific business model and value proposition, not necessarily yours.
• Skipping the validation step and choosing based purely on how a metric sounds conceptually. As covered above, this risks choosing a metric that doesn't actually predict real business success.
• Choosing a metric too complex for the broader organization to understand and rally around. Simplicity is a real requirement, not a nice-to-have.
• Treating the initial choice as permanent, never revisiting it as the product and business evolve. A North Star metric appropriate for an early growth phase may need to evolve as the business matures.

FAQ

How long should the process of defining a North Star metric take? It varies, but rushing this process tends to produce a weaker choice — many teams spend several weeks gathering input, generating candidates, and validating against data before finalizing a North Star metric, rather than deciding in a single meeting.

Who should be involved in choosing a North Star metric? Ideally a cross-functional group, including product, data/analytics, and leadership — since the metric needs both conceptual soundness and genuine organization-wide buy-in to actually function as a shared rallying point.

Can a company change its North Star metric later? Yes, and sometimes it should, as the business and product mature — but changing it too frequently undermines the shared focus and stability the framework is meant to provide, so changes should be deliberate and well-communicated, not frequent.

What if the data doesn't clearly validate any of our candidate metrics? This can be a useful signal that the underlying core value identified in Step 1 needs to be reconsidered, or that more time and data are needed before a confident choice can be validated — proceeding with a best-available hypothesis while continuing to gather validating data is a reasonable interim approach.